Workshop 2

Place: Rouen, France

Date: January 15 - 20, 2018

Attendees:

• Laurent Heutte

• Caroline Petitjean

• Simon Bernard

• Roger Trullo

• George DC Cavalcanti

• Tsang Ing Ren

• Luiz S. Oliveira

• Alceu Britto Jr.

Activities

Monday, January 15, 2018

• Working meetings: visit the Université de Rouen Normandie.

Tuesday, January 16, 2018

• Working meetings: presentation of past and current research activities of NormaSTIC/LITIS and Centro de Informática/UFPE.

Wednesday, January 17, 2018

• Prof. Tsang Ing Ren presented his research interests.

Thursday, January 18, 2018

• Prof. George DC Cavalcanti presented his current research interests mainly focuses on Multiple Classifier System and time series forecasting.

  • Dynamic Classifier Selection

    • Multiple Classifier Systems (MCS) have been widely studied as an alternative for increasing accuracy in pattern recognition. An MCS is essentially composed of three stages: (1) Generation, (2) Selection, and (3) Aggregation orFusion. One of the most promising MCS approaches is Dynamic Selection (DS), in which the base classifiers are selected on the fly, according to each new sample to be classified. This seminar shows some recent advances for the Generation and Selection stages in DS systems.

  • A perturbative approach for time series forecasting

    • We present a method to perform time series prediction based on the perturbation theory. The approach is based on continuously adjusting an initial forecasting model to asymptotically approximatea desired time series model. First, a predictive model generates an initial forecasting for a time series. Second, a residual time series is calculated as the difference between the original time series and theinitial forecasting. If that residual series is not white noise, then it can be used to improve the accuracy of the initial model and a new predictive model is adjusted using residual series. The whole process is repeated until convergence or the residual series becomes white noise. The output of the method is then given by summing up the outputs of all trained predictive models in a perturbative sense.

Friday, January 19, 2018

• Definition of the schedule and program for the next meeting that will take place in Recife-Brazil (November/2018).

Saturday, January 20, 2018

• Trip back home